Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A computing problem management method, the method comprising: launching a plurality of versions of an action, each version being launched and performed on a different replica of a computing system, wherein the each version of the replica includes one of a mistake, a fault, an attack, and a failure for the action to resolve while running the action on the computing system, and wherein each version of the replica includes a different one of the mistake, the fault, the attack, and the failure for the action to resolve, and wherein a number of the plurality of replicas is based on a criticality of resolving an impending problem, further comprising: detecting an impending problem of the computing system; spawning the plurality of replicas when the detecting detects the impending problem; introducing a plurality of new impending problems on the plurality of replicas for the launching to launch versions of the action to resolve; and learning a version that resolves each of the of new impending problems, wherein the launching decides the plurality of versions of the action to launch based on a genetic algorithm including a random selection.
This invention relates to computing problem management, specifically a method for proactively resolving impending system issues by simulating and testing multiple problem scenarios in parallel. The method addresses the challenge of efficiently identifying and mitigating potential system failures, faults, attacks, or mistakes before they impact the primary computing system. The method involves detecting an impending problem in a computing system and then spawning multiple replicas of the system. Each replica is intentionally introduced with a different type of problem—such as a mistake, fault, attack, or failure—to test and resolve. The number of replicas created depends on the criticality of the impending problem, ensuring higher-risk issues receive more extensive testing. Versions of an action are launched on each replica to resolve the introduced problem, with each version designed to address a specific type of issue. The method uses a genetic algorithm with random selection to determine which versions of the action to launch, optimizing the testing process. Additionally, new impending problems are introduced into the replicas to further test and refine the resolution actions. The system learns which versions of the action successfully resolve each problem, improving future problem management. This approach enhances system resilience by proactively identifying and mitigating potential issues through controlled, parallel testing.
2. The method of claim 1 , wherein each replica including a version of the action is presented to a user for selection by the user as to which version to perform, and wherein each replica performs a different version of the plurality of versions of the action to resolve the computing problem.
This invention relates to a system for resolving computing problems by presenting multiple versions of an action to a user, allowing the user to select the most appropriate version for execution. The method involves generating multiple replicas of an action, where each replica contains a distinct version of the action designed to address a specific computing problem. These replicas are then presented to a user, who can evaluate and select the version they believe will best resolve the issue. Once selected, the chosen replica executes its version of the action to address the problem. The system ensures that different versions of the action are tested and applied, increasing the likelihood of finding an effective solution. This approach is particularly useful in scenarios where the optimal solution is uncertain, and user input can help determine the best course of action. The method may be applied in various computing environments, including software troubleshooting, system optimization, and automated problem resolution.
3. The method of claim 1 , further comprising: introducing a plurality of new impending problems on the plurality of replicas for the launching to launch versions of the action to resolve; and learning a version that resolves each of the plurality of new impending problems to thereby create an antifragile computing system, wherein a number of the replicas with the actions launched is based on the cost.
This invention relates to antifragile computing systems that improve through exposure to problems. The technology addresses the challenge of creating systems that not only withstand failures but actively strengthen themselves by learning from and resolving emerging issues. The method involves deploying multiple replicas of a computing system, each capable of executing different versions of an action to address potential problems. New impending problems are intentionally introduced to these replicas, and the system learns which versions of the action effectively resolve these problems. The number of replicas with actions launched is determined by cost considerations, balancing the need for robustness with resource efficiency. By continuously adapting to and resolving problems, the system becomes more resilient and capable of handling future challenges. The approach leverages redundancy and learning to build a self-improving, antifragile computing environment.
4. The method of claim 1 , wherein the action comprises any one of: a software patch; different versions of the software patch; a known solution to the impending problem; a potential solution to the impending problem; and a change in a time of implantation of the action.
This invention relates to proactive software maintenance and problem resolution in computing systems. The technology addresses the challenge of anticipating and mitigating software issues before they cause system failures or performance degradation. The method involves detecting an impending problem in a software system, such as a vulnerability, bug, or performance bottleneck, and automatically implementing a corrective action to prevent the issue from occurring. The corrective actions include applying a software patch, deploying different versions of a patch, implementing a known solution to the problem, testing a potential solution, or adjusting the timing of the action's deployment. The system monitors the software environment for indicators of potential problems, analyzes the data to predict issues, and selects the most appropriate action based on the nature and severity of the impending problem. This proactive approach reduces downtime, enhances system reliability, and minimizes the need for reactive troubleshooting. The invention is particularly useful in large-scale computing environments where manual intervention is impractical or inefficient.
5. The method of claim 1 , wherein the launching decides the plurality of versions of the action to launch based on any of: a genetic algorithm including a random selection; a white noise; a Gaussian noise; a voting; a controlled spread relating to a mean and a standard deviation; and a multidimensional distribution.
This invention relates to a method for launching multiple versions of an action in a system, addressing the challenge of optimizing action selection in dynamic environments. The method determines which versions of an action to launch based on various probabilistic and algorithmic techniques. These techniques include genetic algorithms with random selection, white noise, Gaussian noise, voting mechanisms, controlled spread around a mean and standard deviation, and multidimensional distributions. The genetic algorithm approach uses evolutionary principles to iteratively refine action selection, while noise-based methods introduce variability to explore different outcomes. Voting mechanisms aggregate decisions from multiple sources, and controlled spread ensures diversity within predefined statistical bounds. Multidimensional distributions allow for complex, multi-factor decision-making. The method dynamically adjusts the selection process to adapt to changing conditions, improving system performance and robustness. This approach is particularly useful in applications requiring adaptive behavior, such as autonomous systems, robotics, or decision-making algorithms where optimal actions must be determined under uncertainty.
6. The method of claim 1 , further comprising learning successful versions of the action based on a prior starting state, the version of the action, and a result of the version of the action.
This invention relates to adaptive action selection in automated systems, particularly for improving decision-making based on prior outcomes. The core method involves executing an action in response to a detected state, where the action is selected from a set of possible actions. The system then evaluates the result of the executed action and updates a model to improve future action selection. The improvement includes learning successful versions of the action by analyzing prior starting states, the specific version of the action taken, and the resulting outcome. This learning process refines the system's ability to choose the most effective action for similar states encountered later. The method may also involve tracking multiple versions of an action, comparing their effectiveness, and prioritizing versions that consistently yield better results. The system can be applied in various domains, such as robotics, autonomous systems, or decision-making algorithms, where adaptive behavior based on historical performance is beneficial. The key innovation lies in dynamically refining action selection by leveraging past experiences to enhance future decision-making accuracy.
7. The method of claim 1 , wherein a number, a location on the computing system, and a type of the versions of the action of the plurality of replicas are based on a cognitive state of a user.
This invention relates to a system for managing and executing actions across multiple replicas of a computing system, where the selection and configuration of action versions are dynamically adjusted based on the cognitive state of a user. The system addresses the challenge of optimizing performance and user experience by adapting to the user's cognitive load, attention, or other mental states, ensuring that actions are executed in a manner that aligns with the user's current cognitive capacity. The method involves monitoring the user's cognitive state through sensors, biometric data, or behavioral analysis to determine factors such as focus, stress, or fatigue. Based on this assessment, the system selects a specific number of replicas to engage, determines the optimal location within the computing system where these replicas should operate, and chooses the appropriate version of the action to execute. For example, if the user is highly focused, the system may prioritize speed and efficiency by selecting a minimal number of replicas with high-performance action versions. Conversely, if the user is distracted or fatigued, the system may distribute the action across more replicas with simplified or redundant versions to ensure reliability and reduce cognitive burden. The system dynamically adjusts these parameters in real-time to maintain an optimal balance between performance and user experience, ensuring that actions are executed in a way that minimizes disruptions and maximizes effectiveness. This approach is particularly useful in environments where user attention and cognitive resources are limited, such as in high-stress or multitasking scenarios.
8. The method of claim 1 , wherein, if the detecting detects the impending problem, a processing speed of the computing system is decreased.
A computing system monitors for impending problems such as overheating, power supply issues, or hardware failures. When such a problem is detected, the system reduces its processing speed to mitigate the risk of system failure or damage. This reduction in processing speed may involve throttling the central processing unit (CPU), reducing clock speeds, or limiting the execution of non-critical tasks. The system may also adjust other operational parameters, such as power consumption or thermal management, to further stabilize performance. By dynamically adjusting processing speed in response to detected issues, the system avoids catastrophic failures and extends operational reliability. The method ensures continuous operation under adverse conditions while minimizing the impact on user experience. The system may also log detected problems for diagnostic purposes, allowing for proactive maintenance and troubleshooting. This approach is particularly useful in high-performance computing environments where hardware stress is common, such as data centers, servers, or embedded systems. The solution balances performance and safety, ensuring that the system remains functional even when facing potential hardware or environmental challenges.
9. The method of claim 1 , wherein, if the detecting detects the impending problem, a processing speed of the computing system is decreased such that the plurality of versions of the action are performed by the plurality of replicas at a rate faster than a rate of propagation of the impending problem in the computing system.
A computing system monitors for impending problems, such as hardware failures or software corruption, that could propagate across multiple replicas of a distributed system. The system detects these issues by analyzing performance metrics, error logs, or other indicators. If an impending problem is identified, the system reduces the processing speed of the computing system to ensure that multiple versions of an action are executed by the replicas at a rate faster than the problem's propagation speed. This allows the system to maintain consistency and reliability by completing the actions before the problem spreads, preventing data corruption or service disruptions. The method involves dynamically adjusting the system's performance to prioritize fault tolerance and data integrity over speed when a potential failure is detected. The system may also include mechanisms to revert to normal operation once the problem is resolved or contained. This approach is particularly useful in distributed databases, cloud computing environments, or any system where multiple replicas must remain synchronized despite potential failures.
10. The method of claim 1 , wherein the spawning spawns the plurality of replicas in a cloud-computing system.
This invention relates to cloud computing systems and addresses the challenge of efficiently managing and deploying multiple replicas of a computing resource. The method involves spawning a plurality of replicas in a cloud-computing environment, where each replica is a copy of an original computing resource, such as a virtual machine, container, or application instance. The spawning process ensures that the replicas are distributed across the cloud infrastructure to optimize performance, reliability, and resource utilization. The method may include dynamically adjusting the number of replicas based on demand, load balancing traffic among them, and ensuring high availability by distributing replicas across different geographic regions or availability zones. The cloud-computing system provides the necessary infrastructure, including virtualization, orchestration, and networking capabilities, to support the creation and management of these replicas. The method may also involve monitoring the replicas for performance and health, automatically scaling them up or down as needed, and ensuring seamless failover in case of failures. This approach enhances scalability, fault tolerance, and efficiency in cloud-based applications and services.
11. The method of claim 1 , wherein the spawning spawns the plurality of replicas in an environment with an unlimited resource.
A system and method for managing computational resources in a distributed computing environment addresses the challenge of efficiently allocating and utilizing resources to support multiple replicas of a computational task. The method involves spawning a plurality of replicas in an environment where resources are unlimited, ensuring that each replica operates without constraints on computational power, memory, or other system resources. This approach allows for scalable and flexible execution of tasks, particularly in scenarios where resource demands may vary or where high availability and fault tolerance are critical. The method may include dynamically adjusting the number of replicas based on workload requirements, optimizing resource usage while maintaining performance. The unlimited resource environment ensures that replicas can be instantiated and operated without competition for shared resources, improving reliability and efficiency. This technique is particularly useful in cloud computing, virtualized environments, and distributed systems where resource management is essential for maintaining system stability and performance.
12. The method of claim 1 , wherein the spawning spawns a plurality of children replicas having a relationship to a replica of the plurality of replicas, and wherein each of the plurality of children replicas includes a modified action of the replica with the relationship to the plurality of children replicas.
This invention relates to a method for managing replicas in a distributed computing system, particularly for optimizing task execution by dynamically generating and modifying child replicas. The problem addressed is the inefficiency in traditional systems where replicas perform identical actions, leading to redundant computations and suboptimal resource utilization. The method involves spawning multiple child replicas from a parent replica, where each child replica inherits a modified version of the parent's action. The modifications are based on predefined relationships between the parent and child replicas, allowing for specialized or optimized task execution. For example, a parent replica handling a general computation task may spawn child replicas that perform variations of that task, such as parallelized sub-tasks or optimized subroutines. This approach improves efficiency by tailoring each child replica's behavior to specific requirements, reducing redundancy and enhancing performance. The relationships between replicas can be hierarchical, where child replicas further spawn their own descendants, creating a multi-level structure. Alternatively, relationships may be peer-based, where child replicas operate independently but share modified actions derived from the parent. The modifications can include changes in input parameters, execution logic, or resource allocation, ensuring that each child replica contributes uniquely to the overall task. This method is particularly useful in distributed systems, cloud computing, and parallel processing environments where dynamic adaptation of replicas can significantly improve performance and resource utilization. By leveraging modified actions in child replicas, the system avoids redundant computations and optimizes task
13. The method of claim 1 , wherein the impending problem comprises a system crash.
A system monitoring and predictive maintenance method detects and prevents system crashes by analyzing operational data in real-time. The method involves continuously collecting performance metrics such as CPU usage, memory consumption, disk I/O, and network latency from a computing system. Machine learning algorithms process this data to identify patterns indicative of an impending system crash, such as abnormal spikes in resource utilization or error rates. Upon detecting a high-risk condition, the method triggers automated corrective actions, including resource reallocation, process termination, or system reboot, to mitigate the risk before a crash occurs. The system may also log diagnostic data for post-incident analysis. The method is applicable to servers, embedded systems, and other computing environments where uptime and reliability are critical. By proactively addressing potential failures, the system reduces downtime and improves operational stability. The invention differs from traditional monitoring tools by using predictive analytics rather than reactive thresholds, allowing for earlier intervention and more precise corrective measures.
14. The method of claim 1 , wherein the impending problem comprises any of: an impending device software or hardware status change including a fault, a system crash, a hard-to-recover state, and a speed of the impending device; a state change; and a hardware fault.
This invention relates to predictive monitoring and management of device performance, particularly for identifying and mitigating impending issues before they cause system failures or performance degradation. The method involves detecting early indicators of potential problems in a device, such as software or hardware status changes, system crashes, hard-to-recover states, or abnormal speed variations. It also monitors state changes and hardware faults that could lead to operational disruptions. The system analyzes these indicators to predict and address issues proactively, preventing downtime or performance loss. The method may include collecting real-time data from the device, processing it to identify patterns or anomalies, and triggering corrective actions or alerts based on the analysis. This approach enhances reliability and efficiency in device operation by anticipating and resolving problems before they escalate. The invention is applicable to various devices, including computing systems, industrial equipment, and embedded systems, where predictive maintenance and fault detection are critical.
15. The method of claim 1 , wherein the computing system comprises any of: a computer; a smart phone; a smart watch; a head-mounted display; a game console; and a network component.
This invention relates to a method for processing data in a computing system, addressing the need for efficient and adaptable data handling across diverse computing devices. The method involves receiving input data from one or more sources, analyzing the data to determine its type and relevance, and then processing the data based on predefined rules or machine learning models. The processing may include filtering, transforming, or categorizing the data to optimize performance, storage, or usability. The system dynamically adjusts its operations based on the device's capabilities, such as processing power, memory, or display characteristics, ensuring compatibility and efficiency. The method also supports real-time data processing, allowing for immediate feedback or adjustments. The computing system can be any of various devices, including computers, smartphones, smartwatches, head-mounted displays, game consoles, or network components, enabling broad applicability across different platforms. The invention aims to improve data management by adapting to the specific requirements and constraints of each device, enhancing user experience and system performance.
16. The method of claim 1 , wherein a success of the action is automatically determined if the computing system continues to operate.
A system and method for monitoring and ensuring the success of automated actions in computing systems. The technology addresses the challenge of verifying whether an automated action, such as a software update, configuration change, or system reboot, has been successfully executed without causing system failure. The method involves initiating an automated action on a computing system and then determining the success of the action by monitoring the system's operational status. If the computing system continues to operate normally after the action is performed, the action is automatically deemed successful. This approach eliminates the need for manual verification and reduces the risk of undetected failures. The system may also include additional checks, such as validating system logs or checking for error messages, to further confirm the success of the action. The method is particularly useful in environments where automated actions are frequent, such as cloud computing, data centers, or enterprise IT systems, where ensuring system stability is critical. By automating the success determination process, the system improves efficiency and reliability in managing computing systems.
17. A non-transitory computer-readable recording medium recording a computing problem management program, the program causing a computer to perform: launching a plurality of versions of an action, each version being launched and performed on a different replica of a computing system, wherein the each version of the replica includes one of a mistake, a fault, an attack, and a failure for the action to resolve while running the action on the computing system, and wherein each version of the replica includes a different one of the mistake, the fault, the attack, and the failure for the action to resolve, and wherein a number of the plurality of replicas is based on a criticality of resolving an impending problem, further comprising; detecting an impending problem of the computing system; spawning the plurality of replicas when the detecting detects the impending problem; introducing a plurality of new impending problems on the plurality of replicas for the launching to launch versions of the action to resolve; and learning a version that resolves each of the plurality of new impending problems, wherein the launching decides the plurality of versions of the action to launch based on a genetic algorithm including, a random selection.
This invention relates to a computing problem management system that proactively addresses potential issues in a computing system by simulating and resolving them in parallel replicas. The system detects an impending problem in the computing system and spawns multiple replicas of the system, each containing a different type of problem—such as a mistake, fault, attack, or failure—to be resolved by various versions of an action. The number of replicas generated depends on the criticality of the impending problem, ensuring higher-risk issues receive more extensive testing. Each replica is assigned a unique problem, and different versions of the action are launched across the replicas to determine which version effectively resolves each problem. The system introduces additional new impending problems into the replicas to further test and refine the action versions. A genetic algorithm, incorporating random selection, guides the decision-making process for launching these action versions, optimizing the resolution strategies. The system learns from the outcomes to identify the most effective versions for resolving each problem, enhancing the system's resilience against future issues. This approach allows for dynamic, adaptive problem-solving by leveraging parallel testing and machine learning techniques.
18. A computing problem management system, said system comprising: a processor; and a memory, the memory storing instructions to cause the processor to perform: launching a plurality of versions of an action, each version being launched and performed on a different replica of a computing system, wherein the each version of the replica includes one of a mistake, a fault, an attack, and a failure for the action to resolve while running the action on the computing system, and wherein each version of the replica includes a different one of the mistake, the fault, the attack, and the failure for the action to resolve, and wherein a number of the plurality of replicas is based on a criticality of resolving an impending problem, further comprising: detecting an impending problem of the computing system; spawning the plurality of replicas when the detecting detects the impending problem; introducing a plurality of new impending problems on the plurality of replicas for the launching to launch versions of the action to resolve; and learning a version that resolves each of the plurality of new impending problems, wherein the launching decides the plurality of versions of the action to launch based on a genetic algorithm including a random selection.
This system addresses the challenge of proactively managing computing system problems by simulating and resolving potential issues before they occur. The system detects an impending problem in a computing system and automatically spawns multiple replicas of the system. Each replica is intentionally introduced with a different type of problem, such as a mistake, fault, attack, or failure, to test and resolve the impending issue. The number of replicas created depends on the criticality of the problem, ensuring higher-risk issues receive more extensive testing. The system then launches multiple versions of an action on these replicas, each designed to resolve a specific problem. A genetic algorithm randomly selects which versions of the action to launch, allowing the system to explore various solutions efficiently. The system learns which versions successfully resolve the problems, optimizing future responses. This approach improves system resilience by preemptively identifying and mitigating potential failures through controlled experimentation.
Unknown
May 5, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.